As the data areal density in hard disk drives (HDD) continuously increases because of technology improvements, the MR sensor that is used as the read-back element in HDD is required to have increasingly better special resolution while maintaining a reasonable signal-to-noise ratio (SNR). The sensor is a critical component in which different magnetic states are detected by passing a sense current through the sensor and monitoring a resistance change. A common giant magnetoresistive (GMR) configuration includes two ferromagnetic layers which are separated by a non-magnetic conductive layer in the sensor stack. One of the ferromagnetic layers is a pinned layer wherein the magnetization direction is fixed by exchange coupling with an adjacent anti-ferromagnetic (AFM) pinning layer. The second ferromagnetic layer is a free layer wherein the magnetization vector can rotate in response to external magnetic fields. In the absence of an external magnetic field, the magnetization direction of the free layer is aligned perpendicular to that of the pinned layer by the influence of hard bias layers on opposite sides of the sensor stack. When an external magnetic field, is applied by passing the sensor over a recording medium along an air bearing surface (ABS), the free layer magnetic moment may rotate to a direction which is parallel to that of the pinned layer which is a lower resistance state than in the perpendicular alignment. Alternatively, in a tunneling magnetoresistive (TMR) sensor, the two ferromagnetic layers are separated by a thin non-magnetic dielectric layer.
Referring to FIG. 1a, a generic TMR head 20 is shown and represents the major MR sensor structure that is used in state of the art HDD. A typical TMR head has a top shield 2 and bottom shield 1 with a sensor stack 6 including a free layer (not shown) that is formed between the top and bottom shields and between adjacent hard bias magnets 4. Hard bias structures 4 with a longitudinal magnetization 5 provide a biasing magnetic field on the sides of the sensor stack 6 to orientate the free layer magnetization 12 (FIG. 1b) in the x-axis direction. The thickness of the sensor stack is also referred to as the reader shield spacing (RSS) 3. There is an insulation layer 13 which separates the sensor stack 6 from hard bias structure 4.
In FIG. 1b, a sensor stack 6 having a bottom spin valve configuration is depicted wherein a pinned layer 7, coupling layer 8, reference layer 9, tunnel barrier layer 10, and free layer 11 having a magnetization direction 12 are sequentially formed on the bottom shield (not shown). In addition, there is usually an anti-ferromagnetic (AFM) layer (not shown) between the bottom shield and pinned layer 7, and a top electrode or capping layer (not shown) that separates free layer 11 from top shield 2. Current efforts to further increase area data density involve developing a greater data linear density along a down-track (y-axis) direction and a higher track density along the cross-track (x-axis) direction. Along with a higher track density, a read head with higher spatial resolution in the cross-track direction is required.
Referring to FIG. 2, a schematic drawing of a read-back cross track profile is illustrated which is obtained by scanning the read head across a given data track and plotting the read-back amplitude vs. the off-track distance (distance from track center). The 100% amplitude is the read-back signal when the head is positioned perfectly at track center while uMRW-10% and uMRW-50% are the 10% and 50% micro magnetic read widths that are defined by the width of the cross track profile in FIG. 2 at amplitudes corresponding to 10% and 50% of the track center amplitude. A higher cross track resolution read head means lower uMRW-50% and uMRW-10% values. To achieve higher read-back resolution in the cross track direction, the read head will need to have less side reading of data tracks on the sides of the current track to reduce interference when track density is higher.
Generally, reducing read width is accomplished by reducing the cross track width of sensor stack 6 (width along x-axis in FIG. 1b). In addition, it is known that a smaller RSS 3 and implementing a side shield in the read gap (not shown) can further reduce the effective read width to enhance cross-track resolution. Sensor stack width is effectively the width of free layer 11 which is the distance w.
Referring to FIGS. 3a-3b, simulation results of a read head with different physical sensor stack widths w (FIG. 3c) are shown. The x-axis is effectively uMRW-50% and the y-axis is the ratio (uMRW-10%/uMRW-50%) which represents resolution sharpness. Curve 41 in FIG. 3a is a plot of signal sharpness vs. uMRW-50% for a sensor stack width of 25 nm (data point 41a), 35 nm (data point 41b), and 45 nm (data point 41c) and a RSS 3a of 30 nm (FIG. 3c). Note that sensor stack 6a is essentially equivalent to sensor stack 6 in FIG. 1 except the cross-sectional profile in FIG. 3c is shown as a rectangle rather than a trapezoid and gap layers 16a, 16b are included between sensor stack 6a and top and bottom shields 1, 2, respectively. In this case, RSS 3a is equal to the thickness of sensor stack 6a plus the thickness of gap layers 16a, 16b. Also, the hard bias (HB) structure 4 in FIG. 1 is not shown in FIG. 3c. Curve 42 (FIG. 3a) is similar to curve 41 except RSS 3a is reduced to 20 nm. Therefore, data points 42a, 42b, 42c represent a sensor stack width of 25, 35, and 45 nm, respectively, and a RSS of 20 nm. Curve 41 in FIG. 3a indicates a clear trend of narrower sensor stack width producing a narrower read width (uMRW-50%). More importantly, a smaller RSS as shown in Curve 42 compared with Curve 41 also yields a narrower read width. In FIG. 3b, an additional curve 43 is plotted and represents a condition where side shields 1a, 1b are added in the read gap near the sides of the sensor stack 6a (FIG. 3d) to generate an edge gap 17 with a distance between side shield 2a (or 2b) and free layer in sensor stack 6a of 2 nm. Data points 43a, 43b, 43c represent a sensor stack width of 25, 35, and 45 nm, respectively. Spacing 3a between side shield 2a (or 2b) and bottom shield 1 is reduced to 2 nm. When side shields are present, further reduction in read width is demonstrated. By taking advantage of a smaller RSS (FIG. 3a) or including side shields (FIG. 3b), the ratio (uMRW-10%/uMRW-50%) decreases together with smaller uMRW-50% values thereby indicating reduction in uMRW-10% values as well. It follows that by implementing narrower RSS and side shields, read head cross-track resolution can be improved significantly.
It should be understood by those skilled in the art that reducing RSS 3 and inserting side shields adjacent to sensor stack 6 in FIG. 1a is very difficult to achieve. In particular, a smaller RSS 3 means the thickness of HB structure 4 must decrease accordingly. Unfortunately, a thinner HB structure 4 will lead to a weaker pinning field on edges of free layer 11 and will yield a less stable sensor. Meanwhile, magneto-static coupling between HB structure 4 and top shield 2 will become greater as the spacing between the aforementioned elements decreases which can easily result in a rotation of HB magnetization 5 away from a longitudinal direction in the proximity of free layer 11. Thus, stronger coupling between top shield 2 and HB structure 4 will degrade the effective HB field on the free layer edge magnetization.
With regard to a side shield formed adjacent to a conventional MR sensor, U.S. Pat. No. 6,943,993 describes a structure similar to FIG. 4. Referring to FIG. 4, a portion 50a of top read shield 50 is extended into the read gap (not shown) on either side of a MR sensor stack 46 in an attempt to shield the sensor from being affected by adjacent data tracks (not shown). Spacers 47 separate the MR sensor stack 46 from the HB structure 48. The most obvious drawback of this design is that top shield portion 50a is brought quite close to the HB structure 48 such that HB to top shield coupling will reduce the HB stabilizing field 48a on free layer edges 45 significantly and cause the free layer to be extremely unstable and thereby produce a noisy sensor.
To minimize the HB to top shield coupling issue, perpendicular easy axis growth in a HB structure as disclosed in U.S. Patent Application 2008/0117552 may be employed for a sensor requiring a hard bias structure for stabilization. However, it is even more desirable if a sensor can be designed without a HB structure to avoid the HB-shield coupling entirely.
In addition to the presence of a hard bias structure being a limiting factor in further reducing read gap spacing and sensor track width, the free layer is another contributor to instability at narrow sensor width, especially when a side shield is employed. When the free layer (FL) has an in-plane magnetization and as track width is reduced to accommodate higher track density, the edge surface demagnetization field (demag field) from the magnetic charges created by FL magnetization contacting the confined cross-track edges of the FL becomes stronger. For sensor widths below 30 nm and a read gap dimension less than 20 nm on the side of a sensor stack, it is likely that a HB structure 4 as in FIG. 1 is not able to produce enough stabilizing field to compensate for a build up in FL edge demag field. As a result, the FL edge is not well pinned and undesirable magnetization fluctuations from thermal excitation will occur as described by Y. Zhou in “Thermally Excited Low Frequency Magnetic Noise in CPP structure MR heads”, IEEE Trans. Magn., vol. 43, pp. 2187 (2007). Additionally, when side shields are included in the read head and there is a side shield to FL edge gap of 2 nm or less, strong coupling will arise between charges on the FL edge and on the side shield edge that face each other. This coupling will lead to more free layer instability.
In order to overcome the shortcomings of the prior art and to achieve a high performance MR sensor for both narrow read gap and narrow track width, the following requirements are needed when considering devices with a data area density of greater than 1 Tb/inch2: (1) magnetic biasing is achieved on free layer magnetization without any permanent magnet hard bias structure; (2) magnetic biasing strength is not affected by or can be easily compensated for at narrow read gap distances; (3) free layer self demag field does not produce a higher destabilizing effect on the free layer at smaller design sizes; and (4) magnetostatic coupling from side shield edges on the FL edge is minimized.
In other prior art references, Y. Ding et al. in “Magneto-Resistive Read Sensor with Perpendicular Magnetic Anisotropy”, IEEE Trans. Magn., vol. 41, pp 707 (2005), a magnetic sensor with a free layer having PMA is discussed but is focused primarily on the FL material properties and does not propose any viable sensor structure.
Referring to FIG. 7, U.S. Patent Application Publication 2009/0185315 shows a FL with PMA and a side shield where a top shield 62 and bottom shield 61 are separated by an insulator layer 66. The sensor stack has a reference layer 63, junction layer 64, and free layer 65 formed sequentially on the bottom shield. However, an intrinsic flaw of this design is that FL 65 is deposited on the junction layer which is not feasible with TMR sensors having an oxide junction layer since there must be a magnetic layer abutting the junction layer to realize a high MR signal. When a FL with PMA is formed directly on an oxide layer such as junction layer 64, it is difficult to achieve good PMA. A buffer/seed layer is typically required to establish PMA in an overlying FL but the buffer/seed layer is non-magnetic and cannot produce a high MR ratio as mentioned previously. Therefore, an improved MR read head design is needed that incorporates a FL with high PMA and a TMR sensor configuration without an adjacent hard bias layer for stabilization.
U.S. Pat. No. 7,532,442 teaches a pinning layer made of CoPt in a bottom spin valve or dual spin valve configuration but the CoFe/NiFe free layer does not have any PMA character.
In U.S. Patent Application Publication 2006/0139028, 2-D superparamagnetic bodies are used in a free layer to minimize the free layer thickness needed for a high MR ratio and to avoid a longitudinal biasing structure.